Last week Goldman Sachs Electronic Trading told clients it would begin reporting volume executed in Sigma X, its dark pool, based on single-counted matched executions. This is a departure from how Goldman has been calculating Sigma X volume. Dark pool operators generally are now mulling their options after the Securities and Exchange Commission expressed concern over the lack of reporting standards.
"Without an absolute industry protocol, we’re advocating doing what exchanges have been doing for a very long time: reporting single-counted matched-only volume," said Dave Johnsen, vice president for Sigma X business development at Goldman. "If everyone else puts themselves on the same metric, the dark-pool numbers will add up to 100 percent of dark-pool volume."
Goldman is doing this in part because the SEC recently made noises about wanting more information about how much volume is occurring in dark pools and where that volume is executing. Dark pools are not required to publish this information publicly, although they must print their trades to trade reporting facilities. There is no standard for how dark pools calculate this off-board volume when they issue volume figures.
The SEC’s worry is that this lack of information could impact investors’ execution decisions and impair transparency in the market as dark liquidity grows. SEC Chairman Mary Schapiro said last week that the Commission has heard "concerns from market participants that the lack of post-trade transparency by dark pools makes it difficult, if not impossible, for the public to assess dark pool trading and to identify pools that are most active in particular stocks."
Dark pools accounted for 8.7 percent of consolidated volume in April, compared with roughly 5 percent in early 2008, according to institutional broker Rosenblatt Securities. May figures were not available.
But while Goldman is interested in seeing the industry adopt a uniform reporting methodology, it has not committed to publishing its volume figures publicly every month. Johnsen emphasizes that there should be industry discussion and customer feedback about various alternatives and the degree of information that could be provided publicly, including monthly volumes.
The SEC is particularly concerned about the lack of a uniform reporting methodology for dark pools. James Brigagliano, co-acting director of the SEC’s Division of Trading and Markets, said in a speech last month that dark pools "do not use a uniform methodology and may effectively overstate their true executed volume because of double counting…or by including ‘touched’ volume of orders routed elsewhere for execution with the ‘matched’ volume that they actually execute." Most dark pools double-count their volume, unlike exchanges, which single-count matched trades.
Matched volume is volume executed internally within the ATS by customers. Touched, or handled, volume includes orders routed to other non-displayed venues for executions. That handled volume may be claimed and tallied by more than one firm. "Including handled volume could inflate a dark pool’s numbers," said Adam Sussman, director of research at TABB Group. Goldman’s Johnsen stresses that ATS volume should exclude that handled flow.
Referring to the potential usefulness of "some form of improved post-trade transparency," the SEC’s Brigagliano said that "uniform and reliable trade reporting practices could help establish a fairer playing field because those dark pools that report their volume accurately would not be disadvantaged in comparison with any that might inflate their volume."
Brigagliano did not say how that transparency might be achieved. According to industry participants, one option could be for the SEC or the Financial Industry Regulatory Authority to require ATS venues that do not publish quotes to provide more transparency around their individual prints, identifying them as dark pool trades rather than merely off-board executions printed to a TRF. Another option could be to identify the individual ATS that printed the trade. Regulators could also mandate that ATSs publicly provide aggregate volume statistics, with or without additional information such as the top names traded in each dark pool or the level of price improvement received in those pools. In addition, the industry could voluntarily move toward a standard.
More transparency around executions in dark pools is a good idea, said TABB’s Sussman. He notes that dark pools currently are not obligated to report any volume information publicly. Institutions and brokers, however, need to know how much volume particular pools are executing. "If an institutional client sees that a destination it doesn’t access is increasing its market share, it can connect to that pool or make sure its broker is accessing the destination," Sussman said. "Understanding market share statistics is critical to getting best execution and to ensuring that a firm is accessing the right destinations for its order flow and execution needs."
Two firms that have attempted to provide some transparency around dark pool executions are TABB and Rosenblatt Securities. TABB began publishing its Liquidity Matrix in spring 2007, while Rosenblatt launched "Let There Be Light," its volume breakdown and analysis of dark pool trends in February 2008.
Although many dark pool operators provide some minimal public volume statistics, most provide their clients with much more information. GSET, for instance, gives customers end-of-day reports about their Sigma X and dark pool executions, including execution-quality statistics about their order flow that traded in Sigma X. Johnsen said this information includes the percentage of the spread the customer captured and the amount of price improvement the customer received relative to particular algorithmic benchmarks. This is useful, he said, to traders executing large orders through algos. GSET also provides a real-time account of where customers are getting dark pool executions through Goldman. Sigma X, one of the industry’s largest dark pools, matched a daily average of 123 million shares in May (single-counted).
Credit Suisse, LeveL ATS and other dark pools have no plans to follow Goldman’s lead in how they report their volume, but many firms thinks a single methodology to account for executed volume would be good. Dan Mathisson, head of Advanced Execution Services at Credit Suisse, said his firm isn’t concerned about whether volume is single-counted or double-counted. "We want our numbers to be comparable to the numbers at other dark pools," he said. "Which methodology is used isn’t important, but it’d be nice if everybody didn’t have a different methodology."
Mathisson notes that handled volume should not be included in matched market share reported by dark pools. "It’s relevant to see how much flow is passing through a pool, but handled isn’t matched," he said. CrossFinder matched a daily average of 148 million shares (single-counted) in May, making it the largest pool for a bulge-bracket firm.
One of the industry’s biggest pools is GETCO Execution Services, the only ATS venue that already single-counts its matched executions. GES is a separate broker-dealer owned by the parent of market-making firm GETCO. Launched in March 2008, the ATS executes subscriber flow, including orders from retail brokerages, against flow sent into GES by GETCO’s market-making business.
Knight Link, an electronic dark liquidity product that enables brokers to execute against flow from the market-making division within Knight Capital Markets, single-counts that flow since it tallies only customer executions. The broker-dealer does not report volume for Knight Match, its traditional dark pool. Jamil Nazarali, head of electronic trading at Knight Capital Group, noted that "Knight supports a common standard for reporting trade volumes to make it possible to compare venues."
Whit Conary, president of LeveL ATS, a dark pool owned by five broker-dealers, agrees with Credit Suisse’s approach to a reporting methodology. "The issue isn’t double-counting," he said. "It’s transparency. Customers make routing decisions based on the information they get from venues. That information can’t be confusing and must be accurate." He thinks dark pools "should all be on the same page" in how they report their volume.
Both Credit Suisse and LeveL think real-time identification of where dark pool prints are occurring would be counter-productive. In Mathisson’s view, that type of information could lead to slippage. "I wouldn’t want hedge funds or others to sniff out patterns," he said. Mathisson pointed out that if a fund, for instance, sees three prints in a row at increasing prices in a particular ATS, its models could show that the stock is likely to go up X basis points over the next two days. "If ATS data were released and identified in real time, firms would trawl through the data to find patterns and that could hurt the institutions that use those dark pools," he observed. "More disclosure in that case wouldn’t be a good thing for institutional clients."
Conary, too, sees real-time information as potential information leakage. "Post-trade transparency is critical, but real time almost gets you away from the purpose of being a dark pool," he said. Still, the level of disclosure LeveL is comfortable with may not work for other pools. On its web site, LeveL publishes a list of the highest-volume stocks executed in LeveL on a one-day-delayed basis. The pool publishes a monthly report with aggregate execution-quality statistics that is also publicly available. The majority of other pools don’t make similar information available to the public.
However, even these numbers from LeveL could add some confusion to those looking at the information in a cursory way. The ATS currently double-counts its aggregate volume, in line with the practice at other dark pools, but single-counts its executions in individual stocks on its web site. It does this so traders or quants can make apples-to-apples comparisons of its executed volume in specific names to volume traded in the displayed markets.
With the SEC discussing the importance of post-trade transparency, the issue of what type of post-trade information should be made available, and its frequency, has become an exceedingly thorny subject. Justin Schack, vice president for market structure analysis at Rosenblatt, points out that dark pools may have different views about this based on their business model, their clients and the sensitivity of the flow they execute.
"I think not many people would see a problem with having an identifier that says ‘this print is a dark pool trade,’" Schack said. "But identifying the particular dark pool would be problematic for those pools concerned about information leakage and gaming, particularly if it’s a large print in a thinly traded stock." He added that there’s a broad industry assumption that people should know where liquidity is in the market, including at dark pools. "But if that information is available on a weekly or monthly basis, that’s less valuable than if it were available on a daily basis," Schack said.
Credit Suisse suggests another post-trade reporting option. Regulators, according to Mathisson, could require a centralized reporting database for ATS volume executed by non-displayed markets. He said this could be akin to the program trading stats that exchanges publish.
Mathisson said designating dark-pool volume on a weekly or monthly basis with a time delay "would be healthy." Firms could then see which ATSs have volume, based on a standard reporting methodology. Whether such a database could include additional information about specific stocks or execution quality at dark pools could be discussed by industry participants, he said.
